Project 2 ? Biomarkers of lung cancer risk Project Summary / Abstract The US National Lung Cancer Screening Trial (NLST) demonstrated in 2011 that screening with computed tomography (CT) scans can reduce lung cancer mortality by 20%, but with important costs including a high false-detection rate of 95%. The study also indicated important differences in the benefit of screening in different participant groups as defined by their underlying risk of lung cancer, thus highlighting the urgent need to develop improved risk prediction models for identifying eligible subjects to screen. Biomarker studies of lung cancer have reported a wide range of markers that appear strongly indicative of an individual's underlying risk of developing a cancer that may be combined with information afforded by traditional questionnaire based risk factors, in particular history of tobacco exposure. We hypothesize that a comprehensive and extensively validated risk prediction model that incorporates risk-informative circulating biomarkers has the potential to substantially improve existing risk prediction models, and our pilot data on a limited set of risk biomarkers strongly supports this hypothesis. Our project will focus on systematically assessing a comprehensive panel of biomarkers of lung cancer risk that have been implicated in previous studies, and evaluate the extent to the extent to which they can improve risk prediction. This will be achieved by bringing together data from ongoing large-scale biomarker studies, and conducting a comprehensive de novo analysis of promising risk biomarkers within the Lung Cancer Cohort Consortium (LC3). We will evaluate a wide range of promising risk biomarkers implicated in lung cancer by us or other research groups, including miRNAs, metabolic, immune, and protein biomarkers, as well as epigenetic markers. The initial stage will involve assaying a panel of Promising risk biomarkers at a centralized laboratory for 850 case- control pairs from three prospective cohorts from the US, Europe and Asia, with subsequent large-scale validation of the most informative markers in another 16 prospective cohorts from the LC3 consortium, including 1,750 additional case-control pairs. The versatility of the individual LC3 cohorts will also allow us to thoroughly assess important US minorities, including African American and Hispanic subjects. All lung cancer cases that will be included were diagnosed within 5 years of blood collection. We expect that this will result in establishing a distinct panel of validated and informative risk biomarkers for use in lung cancer risk prediction models, which will be further validated in CT screening studies in collaboration with Project 3.